Hermes YouTube Transcript MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HERMES_YT_DOWNLOAD_DIR | No | Directory where raw downloads are cached | |
| HERMES_YT_WHISPER_MODEL | No | Whisper model size | large-v3 |
| HERMES_YT_NORMALIZED_DIR | No | Directory where normalized WAV files are written | |
| HERMES_YT_WHISPER_DEVICE | No | Device for Whisper (cpu or cuda) | cpu |
| HERMES_YT_TRANSCRIPTS_DIR | No | Directory where Markdown/JSON transcript files are saved | |
| HERMES_YT_WHISPER_BEAM_SIZE | No | Beam size for Whisper decoding | 5 |
| HERMES_YT_WHISPER_COMPUTE_TYPE | No | Compute type for Whisper | int8 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| download_youtube_audioC | Download the best available YouTube audio to a local cache path. |
| transcribe_audioC | Normalize an audio file with ffmpeg and transcribe it with faster-whisper. |
| transcribe_youtubeB | Download a YouTube video, normalize the audio, and transcribe it STT-first. |
| save_transcriptC | Persist a transcript as durable Markdown plus JSON sidecar for Obsidian/Thorn. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 4 tools
Each tool has a distinct purpose: downloading audio, transcribing audio files, transcribing YouTube videos directly, and saving transcripts. No overlap or ambiguity.
All tool names follow a consistent verb_noun pattern in snake_case, such as 'download_youtube_audio' and 'transcribe_audio', making them predictable.
With 4 tools, the set is well-scoped for the goal of generating YouTube transcripts. Each tool serves a necessary step in the pipeline.
Covers the full workflow: download, transcribe, save, plus an end-to-end shortcut. Only minor gap: no tool to retrieve or list existing transcripts.